FLOOD01798

• County: Statewide
• Title: Australian Rainfall and Runoff 1998, Revision of Book VI - Estimation of Large to Extreme Floods
• Date: 11/28/1998
• Prepared By: Rory Nathan, Sinclair Knight Merz
• Floodplain - Doc Type: Educational/Technical/Reference Information

Obtained by application ,
J of rainfall intensity ratios i
K: I
..... I
........... I ,
Obtained by extrapolation
l:J. ~ of regression line
~ o '
- ~ <> .!.-J. I
AEP ---(}l '.
... 1:2000 '.
'. ,
. 1:10DD " i
- +i " ~ ...... .....
. 1:500 I + I .... .....
t , V
. 1:200 I l~ 4- .
.--- ~._~----- - - ____._._L. --
" 1:10D Note: fiDed symbols denote original design values and
+ 1:50 hollow symbols denote interpolated/extrapolated estimates
, . , , ' , , . , , ,

10
Duration (hours)

Figiin. 6 Sc;k8llUltle lllllstndlon of extrapolatil)n of regioAaf design estimMes frQlll kmger
to shorter durations.

....."'" I....

I

estimates appear linear in the log-log domain; note
that extrapolation below 12 hours should not be
attempted as the basic estimates obtained from Book
II Section 1 have been derived for 1. 12 and 72 hour
durations and intermediate durations have been
interpolated.

S. Compute ratios of rainfall depth/intensities for different
AEPs to that of a reference AEP (e.g. the 1 in 100
AEP event) for the lower limit of duration obtained
from the linear relationship of steps 3 and 4.

6. Apply these ratios to the reference AEP design rainfall
depthfJntensities of shorter durations to obtain
corresponding rainfall intensities for AEPs rarer than 1
in 100.

7. Apply areal reduction factors to the point rainfall
estimates to convert to areal values.

A worked example of the above procedure is provided
in Section 622.

I

I

I

I

I

(c) Range of application

In the schematic illustration presented in Figure 6, the
linear relationship used to extrapolate down to shorter
durations is limited to 18 hours, as it is seen that the two
distinctive linear segments for the 1 in 50 AEP and 1 in 100
AEP design values range between 2 and 12 hours, and 18
to 72 hours. In other situations the linear relationship in
step 3 above can be extended to 12 hours. This linear
relationship can also be used to interpolate design rainfalls
for intennediate durations (e.g. 36 hours).

The assumption implicit in step 6 of the procedure is
that the extrapolated frequency curves for all shorter
durations are parallel when plotted on a log-probability
graph. The appropriateness of this assumption for low AEP
design rainfall estimates should be tested by comparison
with corresponding PMP estimates, particularly for
durations shorter than say 6 hours. Judgement is required
to determine how well the linear relationship in the
logarithmic domain holds for durations longer than 72
hours.

I
I

I

I

I

I

I

100

I

.E
E
oS
'"
'in
~ 10
:E

I

I

~
c:
'iij
'"

I

I

1
1

I

..........'" v. ~ ~"U"'cUIVII VI L""YO:;: tv II;;;^Uell'l; r IV....U..

3.7 Seasonal Estimates of Large to Extreme
Rainfalls

(a) Theoretical and practical issues

In some situations, Large to Extreme design rainfall
estimates may be required for specific seasons within the
year (see Section 4.6). While there are a number of design
situations in which it would appear sensible to derive
seasonal estimates, there are a number of practical and
theoretical issues that are not easily resolved. Some of the
issues related to the derivation of seasonal design rainfalls
include:

. Seasonal estimates for Large events. Seasonal
design rain tails for AEPs equal to or more frequent
than 1 in 100 cannot be obtained directly from
information provided in BODk II, Section 1. One
approach is to extract seasonal maxima from rainfall
records at a particular site, and then undertake a
frequency analysis to derive seasonal rainfalls (the
Bureau of Meteorology can provide these estimates if
required). This approach can provide inconsistent
seasonal estimates for the larger events because of
the inherent uncertainties in fitting the tails of the
distribution to observed data, though theoretically this
could be overcome by developing a fitting procedure
that jointly fits all the seasonal distributions. In
addition. the seasonal design rainfalls derived from
the at-site data will need to be adjusted to ensure that
the annual design rainfalls are consistent with the
ARR design estimates (the potential for discrepancies
between at-site frequency estimates and ARR values
is discussed in Section 3.2a).

. Seasonal estimates for Rare events. At present there
are no regional frequency estimates of seasonal
rainfalls known to be available for AEPs rarer than 1 in
100. While there is no theoretical reason why a
procedure such as CRC-FORGE could not be applied
to seasonally censored data from a
hydrometeorologically homogeneous region, the
resources required tD undertake the analysis would be
considerable.

100